Tokenando Indices
Open reference prices for the AI economy
Three daily settled indices, computed from disclosed public sources with a fully published methodology: what a blended million tokens costs per lab (TTPI), what a GPU-hour rents for (TCPI), and the spread between API prices and self-hosted inference (TIS). Settled T+1 at 06:00 UTC; settled values are frozen forever.
Latest settled day 2026-08-25 · methodology v1.2.2 · full methodology →
TTPI — Token Price Index
full series →The volume-weighted blended price per million tokens, per lab: input priced at 75%, output at 25% of traffic, weighted by observed routed volume. Free tiers and self-hosting are excluded. Labs with thin coverage drop for the day — never estimated.
Settled 25 August 2026
$1.92per million tokens, blended+1.48%
TCPI — Compute Price Index
full series →The weighted median USD per GPU-hour across disclosed marketplaces and clouds — an offered-price index, labeled as such. Each provider contributes its median offer; a GPU needs 3+ distinct providers to settle. Interruptible/spot prices live in a separate companion series, never the headline.
Preview 26 August 2026 — unsettled
$2.69per GPU-hour-15.94%
TIS — Inference Spread & Frontier Premium
full series →What a lab charges per million tokens minus what the same tokens cost to serve yourself on rented GPUs with an open-weight model. Throughput assumptions come from cited public benchmarks and publish only after human verification. The Frontier Premium is the headline ratio: cheapest closed-lab price over cheapest credible self-hosted equivalent.
Settled 25 August 2026
$0.024spread, $ per million tokens-72.93%
Latest settled values
| Series | Key | Value | Δ day | Coverage |
|---|---|---|---|---|
| TTPI | Anthropic | $8.17/MTok | -0.58% | 16 models · 100% volume |
| TTPI | Moonshot AI | $5.15/MTok | +0.61% | 7 models · 100% volume |
| TTPI | xAI | $2.81/MTok | 0.00% | 6 models · 100% volume |
| TTPI | OpenAI | $1.92/MTok | +1.48% | 57 models · 98% volume |
| TTPI | Z.ai (Zhipu) | $1.86/MTok | +18.80% | 13 models · 100% volume |
| TTPI | Qwen | $1.15/MTok | -0.74% | 50 models · 93% volume |
| TTPI | $0.924/MTok | -1.92% | 24 models · 99% volume | |
| TTPI | MiniMax | $0.523/MTok | -0.06% | 8 models · 100% volume |
| TTPI | Tencent | $0.231/MTok | 0.00% | 6 models · 100% volume |
| TTPI | DeepSeek | $0.209/MTok | -21.73% | 14 models · 100% volume |
| TTPI | Mistral | $0.145/MTok | -0.07% | 20 models · 98% volume |
| TTPI | Meta | $0.141/MTok | -3.61% | 8 models · 100% volume |
| TCPI | B200 | $6.79/GPU-hr | 0.00% | 3 providers · 140 offers |
| TCPI | H200 | $4.08/GPU-hr | +1.91% | 5 providers · 119 offers |
| TCPI | H100 SXM | $3.20/GPU-hr | 0.00% | 6 providers · 169 offers |
| TCPI | H100 PCIe | $2.73/GPU-hr | 0.00% | 9 providers · 113 offers |
| TCPI | A100 SXM4 | $1.39/GPU-hr | -0.71% | 6 providers · 201 offers |
| TCPI | RTX 5090 | $0.650/GPU-hr | 0.00% | 4 providers · 274 offers |
| TCPI | RTX 4090 | $0.397/GPU-hr | +8.11% | 4 providers · 271 offers |
Free JSON API, no key: GET /api/v1/ttpi?lab=anthropic · GET /api/v1/tcpi?gpu=H100%20SXM · endpoints documented in the methodology.
Tokenando indices are informational reference data, not financial products, price quotes, or investment advice. Offered prices are labeled offered; sources, weights, and exclusion rules are fully disclosed in the methodology. Settled values are immutable; corrections ship as new methodology versions, never as silent edits.